Referencing Tool

During scientific research a huge amount of related work is searched. These papers are related to each other by their mutal referencing. Consequently the papers can be seen as a graph, which can be filtered by the actual research direction to interesting subgraphs of works. Through an additional extension it should be possible to recommend not yet found works for example by the number of refrencing papers.
The goal of this thesis is to create a first prototype of this application, capable of creating the graph structure out of given pdf files and saving it to a relational database. On base of this data a visualisation and appropriate processing alogrithm should be implemented to find new related literature.